A research team led by Lawrence Livermore National Laboratory (LLNL) has identified several existing medications potentially associated with improved survival rates for patients with amyotrophic lateral sclerosis (ALS). Published in The Lancet Digital Health, the study leverages artificial intelligence and causal inference techniques to analyze a massive dataset of electronic health records. This collaborative effort involved LLNL, Stanford University School of Medicine, the Veterans Affairs Palo Alto Health Care System, and the University of California, Los Angeles (UCLA). The research examined health records from over 11,000 U.S. military veterans diagnosed with ALS between 2009 and 2019. This specific cohort was selected due to a unique convergence of factors: in 2009, ALS was formally recognized as a service-connected disease, leading to a decade of detailed, centralized treatment data within the Veterans Health Administration. This availability was complemented by new targeted funding programs, enabling large-scale research on a rare disease that has historically lacked sufficient resources. The study was further motivated by recent setbacks in standard drug development, notably the withdrawal of Relyvrio from the market in 2024 after follow-up trials failed to confirm its benefits. Rather than relying on traditional machine learning approaches, the team utilized causal inference methods. This framework is designed to isolate treatment effects and account for biases, confounding factors, and uneven treatment patterns found in real-world data. Priyadip Ray, LLNL principal investigator, noted that the VA system's extensive experience provided an alternative pathway for discovery. The analysis evaluated 162 medications prescribed for other conditions and identified 27 drugs associated with statistically significant changes in mortality risk. Notably, multiple medications within the same therapeutic classes, including statins, phosphodiesterase type 5 inhibitors, and alpha-adrenergic antagonists, demonstrated similar associations with prolonged survival. Ray stated that observing this consistent pattern across drug groups provided strong confidence in the link between these treatments and slowed ALS progression. To understand the biological mechanisms, the team used PathFX, a protein-protein interaction modeling tool developed by the UCLA and Stanford collaborators. The network analysis suggested that these drugs converge on shared downstream protein pathways, highlighting potential common mechanisms and new molecular targets for future research. While the findings do not prove clinical efficacy, they establish a robust foundation for future investigations. The researchers plan to conduct deeper modeling that accounts for time-varying health factors and seek validation in independent datasets that include more diverse civilian populations. To facilitate further research despite the challenges of sharing sensitive medical data, the team is releasing their software pipeline as open source. This allows researchers globally to apply these tools to their own datasets, diseases, and interventions, potentially accelerating the discovery of new therapies for ALS.